AI-power Automation Engineer

AI-POWERED AUTOMATION ENGINEER 

 

We make caps. As in, the baseball cap on your head right now is probably from a factory like ours. 

We also have a budget that most “tech startups” would quietly envy, and we're putting it into one thing: rebuilding how a 1,000-person manufacturing company actually runs, using AI agents. 

That's this role. Below is what you'd actually do, not corporate fluff. 

What’s on your plate (real things from this month): 

1. Clean up our NAS. 
Thousands of files, no naming convention, half of them lost.You read the whole tree with AI, propose a folder structure that fits a garment company, get buy-in from each team, then mass-rename and restructure with scripts. One project, done. 

2. Stitch Sales, Material, and Production data together. 
Today they're three islands. Different IDs, different formats, different rules. You hunt the joins, fill gaps with AI, and produce a master file the GD can actually trust. When you don't know enough about garment ops, you ask AI smart questions until you do. 

3. Prep for ERP migration. 
Clean datasets in, batch-load formats out. When the ERP lands next year, the data is ready and you've already written the migration scripts. 

4. Translate for the field. 
Sales, Material, Production staff bring AI ideas every week. They know the work; they don't know how to ship code. 
You sit with them for 30 minutes and come back two days later with a working tool. 

5. Hunt buyers like an analyst. 
We already use AI to pull competitor filings, websites, disclosures, and back into “who their real customers are.” 
You keep building this muscle. Your output goes straight to the sales team and directly shapes which buyers we land. 

6. Kill the small painful stuff. 
Example: Finance gets daily cash in/out from each subsidiary and manually consolidates it. That's a script. 
There are dozens like it. You find them and kill them. 

Why this isn’t a “career graveyard” role: 

Most devs who land at a garment company end up maintaining 10-year-old ERP screens and stop learning. We built this role to be the opposite. 

• Ship weekly, not quarterly. Two-day prototype, one-week deploy, real users by Friday. That’s the rhythm. 

• Report to the GD. Your work is visible at the top within hours. Politics: close to zero. 

• Define your own stack. Greenfield. Pick the AI agents you actually want to live in. 

• Real impact beats toy projects. A weekend side-project gets 10 users on a good day. Here, one script saves dozens of people an hour a day. That trade is the whole point. 

 

Salary 

• Meaningfully above garment-industry baseline. 

• We treat this seat as a strategic investment, not a back-office hire. 

• If you're good, the offer reflects it. Negotiate. 

 

You’d fit if you: 

• Have shipped at least one real end-to-end thing using an AI coding agent (Claude Code, Cursor, Codex, Windsurf, Copilot, v0, Bolt, Lovable, Replit Agent. Pick your poison.) 
  → Live link beats polished resume. 

• Are comfortable starting from a messy Excel someone emailed you with no docs. 

• When you don't know something, you ask AI better questions instead of waiting for a senior to explain. 

• Hold a working conversation in English or Korean. 

 

Nice to have, not required: 

• SQL 

• Python 

• Manufacturing exposure 

• ERP experience 

• Korean language 

 

 

---

Your Safety and Data Security Matter to Us

Manpower is committed to a safe, secure, and transparent hiring process. We will never request any form of payment, banking details, or sensitive personal information at any stage of recruitment.

If you receive any suspicious communication claiming to be from Manpower, please report it immediately to: [email protected]

---

An toàn & Bảo mật Dữ liệu của bạn là ưu tiên của chúng tôi

Manpower cam kết mang đến quy trình tuyển dụng an toàn, minh bạch. Chúng tôi sẽ không bao giờ yêu cầu bất kỳ khoản phí nào, thông tin tài khoản ngân hàng hoặc dữ liệu cá nhân nhạy cảm trong suốt quá trình tuyển dụng.

Nếu bạn nhận được liên hệ đáng ngờ tự xưng là Manpower, vui lòng báo cáo ngay qua email: [email protected]